We propose to conduct a multi-therapist, prospective, observational study of the outcomes of usual low vision rehabilitation (LVR) services. The purpose of this project is to study the concordance of three different types of LVR outcome measures (patient-reported, therapist-reported, and functional performance), study the influence of therapists' biases on estimated outcome measures, study the effects of co-morbidities and other patient states on LVR outcomes, and study the relationship of the effectiveness of LVR to its efficacy. We aim to recruit 10 LVR therapists as clinical collaborators. Each therapist will be expected to recruit 45 patients over a period of 4 years or less. The therapists will provide their usual low vision rehabilitation services between baseline and follow-up outcome measures. Research assistants will interview patients by telephone and administer the adaptive Activity Inventory (AI), which requires patients to rate the importance and difficulty of activity goals and rate the difficulty of subsidiary tasks under each important ad difficult goal, before and after LVR. The therapists will generate Goal Attainment Scale (GAS) ratings of patients' activity goals and administer a battery of functional performance tests before and after LVR. We have developed a measurement model that will be used to estimate interval-scaled measures of patients' functional abilities in different functional domains from patients' difficulty ratings, therapists' GAS ratings, and functional performance measures. Changes after LVR in patient-reported functional ability measures estimated from patients' difficulty ratings of AI goals will be compared to changes after LVR in functional ability measures estimated from therapists' ratings of the same AI goals using the GAS method. We also will compare changes after LVR in patient-reported measures of functional ability (estimated from patients' difficulty ratings of the appropriate subset of AI tasks) to changes after LVR in corresponding performance measure estimates of functional ability (reading speed versus print size measures for reading function, walking speed measures for mobility function, measures of maze tracing errors for visual motor function, and measures of facial expression recognition probability correct for visual information processing function). The objective of the proposed research is to understand sources of bias in, and effects of confounding variables on, the three different types of LVR outcome measures that can be used in clinical research on LVR.